Fully distributed strain-based output feedback for enhanced sensitivity in damage diagnosis of structures

Abstract

The closed-loop damage diagnosis method has shown significant improvement in recent years, while closed-loop eigenvectors are seldomly used in previous damage diagnosis research due to the lack of adequate measurement points, by which a better damage diagnosis effect can be achieved. With the maturity of fully distributed optical fiber sensing technology, strain data acquisition with many (high-density) measurement points can be realized; in this manner, fully distributed strain-based output feedback for sensitivity enhancement is first investigated in this study, while the main contribution is to first develop a calculation algorithm of closed-loop strain modal shape sensitivity with high-density strain measurements. First, the desired eigenstructure assignment scheme is constructed, and a reasonable range of closed-loop pole selection is determined. On this basis, a sensitivity-based damage diagnosis method using closed-loop strain modal shapes is proposed, with sensitivity and robustness selected as the optimization goals. The optimal gain is obtained, which can benefit the closed-loop damage identification accuracy and stability. The effectiveness of the proposed method is validated and compared with the closed-loop eigenvalue-based method by numerical simulation, and the results indicate that the proposed method can effectively enhance the performance of damage diagnosis and improve system robustness.